Welcome to EDAboard.com

Welcome to our site! EDAboard.com is an international Electronic Discussion Forum focused on EDA software, circuits, schematics, books, theory, papers, asic, pld, 8051, DSP, Network, RF, Analog Design, PCB, Service Manuals... and a whole lot more! To participate you need to register. Registration is free. Click here to register now.

Register Log in

Bandbap refererence issue

Status
Not open for further replies.

dennislau

Member level 1
Joined
Dec 16, 2005
Messages
34
Helped
3
Reputation
6
Reaction score
2
Trophy points
1,288
Activity points
1,638
To design a bandgap reference with standard CMOS technology, basically there are mainly two types of bandgap reference (with paraisitic bipolar transistor), one is use the op-amp type while another is use the current mirror/cascode mirror current mirror (without op-amp) type to generate the temperature, and supply-independent voltage. Which one is better against the process variation? And, does anyone suggest some topology which has good performance against the process variation or with accurate initial bandgap voltage without trimming?
 

sunking

Advanced Member level 3
Joined
May 25, 2004
Messages
874
Helped
70
Reputation
140
Reaction score
23
Trophy points
1,298
Activity points
6,283
with op is better.
both of them need trimming for product.
 

Chethan

Full Member level 3
Joined
Jul 11, 2005
Messages
173
Helped
20
Reputation
40
Reaction score
12
Trophy points
1,298
Location
Bangalore
Activity points
2,841
sunking said:
with op is better.
both of them need trimming for product.
Hi sunking,

can u pls tell which architecture has better PSRR. Does both of them need an output capacitor to reduce the ripple.If they need it then what should be the value of the capacitor.
 

xerxer

Junior Member level 3
Joined
Nov 15, 2005
Messages
29
Helped
0
Reputation
0
Reaction score
0
Trophy points
1,281
Activity points
1,523
i think two structures need to be rippled because the rise of ripples is mainly load ,not of the ref structure.
 

jluhzw

Member level 5
Joined
Nov 15, 2004
Messages
94
Helped
17
Reputation
34
Reaction score
4
Trophy points
1,288
Location
Hangzhou China
Activity points
759
I think the one which have the opamp is better
If you design a good opamp you will attain a better bandgap reference.Beacuse the opamp usually have higher PSRR than the other.
you can turn to the simulation for the capacitor.
 

sathaiah

Newbie level 1
Joined
Jan 6, 2006
Messages
1
Helped
0
Reputation
0
Reaction score
0
Trophy points
1,281
Activity points
1,292
The Op-amp having higher loop gain can give better PSRR performance, but the trade-off is with stability,
i would suggest that, you can include substractor circuit on core, that largly sustracts the ripple on supply, you could see arond 20dB better PSRR spec than the normal architecture. please see the attached PDF for implemetation of sustractor circuit in Bang-Gap.
 

allan_guo

Junior Member level 3
Joined
Jan 6, 2006
Messages
28
Helped
4
Reputation
8
Reaction score
2
Trophy points
1,283
Activity points
1,555
I think:OP is better,but the area of the op is fairly large in order to maintain the stability.Cascode structure require high voltage.if you use wide swing , the PSRR will become worse.
 

kenlino

Junior Member level 2
Joined
Dec 30, 2005
Messages
22
Helped
2
Reputation
4
Reaction score
1
Trophy points
1,283
Activity points
1,398
process variation is the nodus for precise bandgap voltage without trimming
 

dennislau

Member level 1
Joined
Dec 16, 2005
Messages
34
Helped
3
Reputation
6
Reaction score
2
Trophy points
1,288
Activity points
1,638
kenlino said:
process variation is the nodus for precise bandgap voltage without trimming
Yes, I agree. It is difficult to achieve that.....someone propose using the switching technique to do the averaging against the process variation but I think this technique is noisy.
 

haemun

Newbie level 4
Joined
Nov 14, 2005
Messages
7
Helped
0
Reputation
0
Reaction score
0
Trophy points
1,281
Activity points
1,331
Which process do all you mostly use?
 

holddreams

Full Member level 6
Joined
Aug 2, 2005
Messages
351
Helped
15
Reputation
30
Reaction score
7
Trophy points
1,298
Location
Shanghai
Activity points
4,243
How will you start to design the bandgap circuit?

How to determin the two resistors used in bandgap circuit?

I run a .tran simulation ,and at the same time sweep the temperature, but the bandgap output is decreasing as the temperature increase, How to solve this problem?Does it mean that the two resistors I used are wrong?

Thx.
 

dennislau

Member level 1
Joined
Dec 16, 2005
Messages
34
Helped
3
Reputation
6
Reaction score
2
Trophy points
1,288
Activity points
1,638
holddreams said:
How will you start to design the bandgap circuit?

How to determin the two resistors used in bandgap circuit?

I run a .tran simulation ,and at the same time sweep the temperature, but the bandgap output is decreasing as the temperature increase, How to solve this problem?Does it mean that the two resistors I used are wrong?

Thx.
Usually, in the conventional bandgap circuit design, we can use the equation Vref=Vbe+ (kT/q) x (lnN) x (R2/R1) to calculate the two resisors value where N typically chosen to be 8 or 24 from matching point of view .

It is mentioned that the bandgap output is decreasing as the temperature increase. Could you tell us how much it decrease?
 

Chethan

Full Member level 3
Joined
Jul 11, 2005
Messages
173
Helped
20
Reputation
40
Reaction score
12
Trophy points
1,298
Location
Bangalore
Activity points
2,841
dennislau said:
holddreams said:
How will you start to design the bandgap circuit?

How to determin the two resistors used in bandgap circuit?

I run a .tran simulation ,and at the same time sweep the temperature, but the bandgap output is decreasing as the temperature increase, How to solve this problem?Does it mean that the two resistors I used are wrong?

Thx.
Usually, in the conventional bandgap circuit design, we can use the equation Vref=Vbe+ (kT/q) x (lnN) x (R2/R1) to calculate the two resisors value where N typically chosen to be 8 or 24 from matching point of view .

It is mentioned that the bandgap output is decreasing as the temperature increase. Could you tell us how much it decrease?

Hi,
The BGR i have designed has the temperature plot as shown in the figure. The reference voltage drastically decreases as the temperature increases. can some one please tell me how should i compensate it.

thanx in advance
 

dennislau

Member level 1
Joined
Dec 16, 2005
Messages
34
Helped
3
Reputation
6
Reaction score
2
Trophy points
1,288
Activity points
1,638
Chethan said:
dennislau said:
holddreams said:
How will you start to design the bandgap circuit?

How to determin the two resistors used in bandgap circuit?

I run a .tran simulation ,and at the same time sweep the temperature, but the bandgap output is decreasing as the temperature increase, How to solve this problem?Does it mean that the two resistors I used are wrong?

Thx.
Usually, in the conventional bandgap circuit design, we can use the equation Vref=Vbe+ (kT/q) x (lnN) x (R2/R1) to calculate the two resisors value where N typically chosen to be 8 or 24 from matching point of view .

It is mentioned that the bandgap output is decreasing as the temperature increase. Could you tell us how much it decrease?

Hi,
The BGR i have designed has the temperature plot as shown in the figure. The reference voltage drastically decreases as the temperature increases. can some one please tell me how should i compensate it.

thanx in advance

The figure shows that the bandgap circuit works well. Do you mean the temperature coefficient of the bandgap circuit is not good enough as I see the voltage drop around 15mV across the temperature sweep?
 

Chethan

Full Member level 3
Joined
Jul 11, 2005
Messages
173
Helped
20
Reputation
40
Reaction score
12
Trophy points
1,298
Location
Bangalore
Activity points
2,841
yes,

the reference voltage reduces considerably as temperature increases. can someone explain how to correct it.

thanx in advance
chethan
 

allan_guo

Junior Member level 3
Joined
Jan 6, 2006
Messages
28
Helped
4
Reputation
8
Reaction score
2
Trophy points
1,283
Activity points
1,555
Chethan said:
yes,

the reference voltage reduces considerably as temperature increases. can someone explain how to correct it.

thanx in advance
chethan
the curve is correct,the voltage drop with the temp increase because of 2nd effect,
the zero temp coef only one point(usully room temp). Few papers write about the rectify of 2nd curve.
 

jcpu

Full Member level 4
Joined
Dec 17, 2005
Messages
215
Helped
18
Reputation
36
Reaction score
6
Trophy points
1,298
Activity points
2,968
Dear Sir:

Just curious.
Allow me to ask:
If 15mV variation against temp is too much,
then how much do you expect against process variation,
or supply voltage variation?

Have a nice day.
 

dennislau

Member level 1
Joined
Dec 16, 2005
Messages
34
Helped
3
Reputation
6
Reaction score
2
Trophy points
1,288
Activity points
1,638
allan_guo said:
Chethan said:
yes,

the reference voltage reduces considerably as temperature increases. can someone explain how to correct it.

thanx in advance
chethan
the curve is correct,the voltage drop with the temp increase because of 2nd effect,
the zero temp coef only one point(usully room temp). Few papers write about the rectify of 2nd curve.
yes, the TC of bandgap can be further improved by the 2nd order compensation method. Here is one for your information, you can also serach "curvature-compensation" for this issue as there are many different way to implement the curvature-compensation technique.
 

Areky_qin

Advanced Member level 4
Joined
Jan 29, 2005
Messages
100
Helped
8
Reputation
16
Reaction score
3
Trophy points
1,298
Activity points
598
BTY, we have designed a bandgap with a op, the simulation result is very good, just as the power supply range, temperature characteristic, but when we test the chip,the bandgap block not work, so i think there r more things to think about, not only some paper can show all the information, tape-out and test results show all.
 

Status
Not open for further replies.
Toggle Sidebar

Part and Inventory Search

Welcome to EDABoard.com

Sponsor

Top